You are on page 1of 15

Page 1

50 U. Chi. L. Rev. 800, *

1 of 1 DOCUMENT

Copyright © 1983 University of Chicago.


University of Chicago Law Review

SPRING, 1983

50 U. Chi. L. Rev. 800

LENGTH: 10687 words

Statutory Interpretation -- in the Classroom and in the Courtroom

NAME: Richard A. Posner +

BIO:

+ Judge, U.S. Court of Appeals for the Seventh Circuit; Senior Lecturer, University of Chicago Law School. The helpful comments of
Guido Calabresi, Michael Carroll, Frank Easterbrook, Peter Edelman, Philip Elman, Henry Friendly, Susan Webb Hammond, Edward Levi,
Bernard Meltzer, Antonin Scalia, and Cass Sunstein on a previous draft of this paper are gratefully acknowledged, as is the research assist-
ance of James Talent.

SUMMARY:
... You can pick up a clue to the situation by perusing the few published law-school textbooks devoted entirely to legis-
lation. ... I have sketched the elements of a proposed second-year course in law school that, so far as I know, is offered
nowhere today (I would be very happy to learn otherwise). ... There is an initial question of what precisely it means to
call a canon of statutory construction "wrong." ... I do not think that any of the canons of statutory construction can be
defended on the theory that they are keys to deciphering a code. ... And, as noted earlier, two inconsistent canons can
usually be found for any specific question of statutory construction. ... My last example of a canon apparently premised
on an assumption of legislative omniscience is one that even Judge Friendly, our most trenchant living critic of the can-
ons of statutory construction, has occasionally, though cautiously, invoked: that the reenactment without change of a
statute that the courts have interpreted in a particular way may be taken as evidence that the reenactment adopts that
construction. ...

TEXT:
[*800] This paper continues a discussion begun in an earlier paper in this journal. n1 That paper dealt primarily with
the implications for statutory interpretation of the interest-group theory of legislation, recently revivified by economists;
it also dealt with constitutional interpretation. This paper focuses on two topics omitted in the earlier one: the need for
better instruction in legislation in the law schools and the vacuity of the standard guideposts to reading statutes -- the
"canons of construction." The topics turn out to be related. The last part of the paper contains a positive proposal on
how to interpret statutes.
I. THE ACADEMIC STUDY OF LEGISLATION
It has been almost fifty years since James Landis complained that academic lawyers did not study legislation in a
scientific (i.e., rigorous, systematic) spirit, n2 and the situation is unchanged. There are countless studies, many of high
distinction, of particular statutes, but they are not guided by any overall theory of legislation, and most academic law-
yers, like most judges and practicing lawyers, would consider it otiose, impractical, and pretentious to try to develop
one. No one has ever done for legislation what Holmes did for the common law. n3 The closest thing may be the eco-
nomists' version of the interest-group theory, n4 but the economists have limited their attention to a tiny subset of stat-
utes, and their work has [*801] barely begun to penetrate the thinking of mainstream academic lawyers. n5 Although
there is a fine literature debunking the canons of construction, n6 one has only to skim any recent volume of the Federal
Reporter or the United States Reports to discovery that it has had little impact on the judicial reading of statutes. The
enormous political science literature on Congress and the state legislatures, n7 which might cast light on the question
whether it is realistic to ascribe to the draftsmen of legislation a knowledge of the code the courts use to interpret stat-
Page 2
50 U. Chi. L. Rev. 800, *

utes, is unknown at the practical level of the legal profession; so is the older political science literature on the role of in-
terest groups in legislation. n8
While many academic lawyers are experts on particular statutes -- which largely means experts on what the courts
have said about the particular statutes they teach -- few are experts on legislation. Few study legislation as an object of
systematic inquiry comparable to the common law; few attempt even to translate the studies of economists and political
scientists into language that lawyers can understand. You can pick up a clue to the situation by perusing the few pub-
lished law-school textbooks devoted entirely to legislation. n9 None recognizes the fundamental role of interest groups
in procuring legislation, though the existence of such groups is acknowledged, particularly in discussions of legal regu-
lations of lobbying; n10 one text omits statutory construction as a [*802] topic; n11 there is little reference to the eco-
nomic or political science literature on legislation. n12 Most give disproportionate attention to such specialized topics as
reapportionment and the regulation of campaign financing, presumably because these are areas where cases can be
found for a casebook treatment of legislation. This shows that the editors do not really conceive of legislation as a dis-
tinct subject. A casebook on commercial law makes good sense; a casebook on legislation does not.
Nor is it an adequate reply to my Jeremiad that every good course in a statutory field such as commercial law or
taxation or antitrust or copyright will impart to the students, in proper law school inductive fashion, a feel for the recur-
rent issues and problems involving legislation in general. Most teachers of statutory fields believe they have only
enough time to introduce the students to the field -- to give the students a sense of the field's scope and texture by work-
ing through the major statutory provisions and the principal cases construing them. They do not feel they have enough
time to explore with the class the process by which the legislation is enacted, the political and economic forces that
shaped it, or even the methods the courts use to interpret it, as distinct from the particular interpretations that the courts
have made. Moreover, such issues are rarely dealt with in casebooks, and what is not in the casebook is unlikely to get
into the course in any systematic fashion.
There are exceptions to these generalizations about course coverage; I am aware that seminars in legislation are
offered in many law schools (the University of Chicago Law School, for example) and that courses in legal process con-
sider legislation. But the sum of these various, rather fragmentary offerings is meager.
Twenty years after I graduated from law school, law school students, faculty, and administrators continue to com-
plain about the malaise of the second and third years of law school. What better, if extremely limited, answer than a
good course on legislation? The absence of suitable materials is only a short-term impediment.
I shall indicate briefly what I think such a course would contain, in the hope that people who have more time than I
to devote [*803] to curricular innovation will carry these ideas to fruition. n13
1. The Process of Legislation. Many law students are ignorant of the process by which bills in Congress become
law (more do not know even the first thing about the legislative process at the state level). Many law students do not
know who writes a bill, who testifies at hearings, what a conference report is, or the difference between adding an
amendment to a bill on the floor and processing a bill in the usual way through committees. It is as if in reading judicial
opinions they did not know how a case got into an appellate court. Reluctant as law professors are to impart mere in-
formation in the classroom, I can see no escape from their doing so (not necessarily orally) in this instance.
2. The Empirical Study of Legislation. It is not enough simply to give students an outline of how the legislative
process operates and a synoptic view of the scholarly controversy over the nature of what it produces. They should also
be exposed to the results of empirical studies. They ought to learn what political scientists have discovered about the re-
spective roles of Congressmen and staff n14 in drafting legislation, the contribution of lobbyists and administration offi-
cials, and the time and care devoted to actual drafting, so that they can form their own judgment on whether it is realist-
ic to suppose that statutes are drafted in the light of assumptions concerning the methods that courts will use to interpret
them. n15 They should also learn about the frequency and feasibility of legislative overruling of judicial decisions that
interpret statutes contrary to the purpose of the legislation as conceived by either the enacting Congress or a subsequent
one. n16 Such data are essential to understanding and evaluating the judicial role in statutory interpretation, though I am
quick to add that, to my knowledge at least, adequate data have not been compiled -- further evidence of the absence of
a scientific spirit from the study of legislation.
3. Techniques for Judicial Interpretation of Statutes. The student should be introduced to the debunking literature
on the [*804] canons of construction, n17 to the positive literature that is worth reading, n18 to the canons themselves, n19
and to those masterpieces of statutory interpretation, such as Judge Learned Hand's opinion in the Fishgold case. n20
The student will not encounter these works in the regular curriculum because they deal with statutes, such as the veter-
Page 3
50 U. Chi. L. Rev. 800, *

ans' reemployment provision n21


construed in Fishgold, that are not the subject of any regular course. n22

4. Researching Legislative History. A year and a half of reading briefs in cases that often involve statutory inter-
pretation has convinced me that many lawyers do not research legislative history as carefully as they research case law.
They may not know how. It is more difficult to research legislative history than case law, yet instruction in the former
is, at most law schools anyway, redimentary. Often it is crammed into the whirlwind tour of the library that law librari-
ans offer to beginning students who cannot comprehend the significance of what they are being told and shown. The
many sources of compiled legislative histories n23 remain largely unknown to the profession. Research into U.S. gov-
ernment documents in general and legislative documents in particular is a formidable subspecialty of library science, n24
and I would guess that not one lawyer in a thousand has a real proficiency in it. He will not pick it up in his law firm
and he will not learn -- not well anyway -- by doing. The mastery of research techniques is not so intellectually stimu-
lating as other elements of a law school education, [*805] but library science is a recognized field of learning at first-
class universities, and it would not demean the law schools to offer formal instruction in a highly relevant aspect of it.
I have sketched the elements of a proposed second-year course in law school that, so far as I know, is offered
nowhere today (I would be very happy to learn otherwise). New I shall discuss some of the consequences of this gap in
legal education.
II. THE CANONS OF CONSTRUCTION

A. Introduction
The canons of statutory construction -- for example, one starts with the language of the statute; repeals by implica-
tion are not favored; penal statutes are to be construed narrowly and remedial statutes broadly; expressio unius est ex-
clusio alterius n25 -- occupy a kind of legal demimonde. To exaggerate slightly, it has been many years since any legal
scholar had a good word to say about any but one or two of the canons, but scholarly opinion -- and I include not just
the views of professors but the view expressed in nonjudicial writings of distinguished judges such as Frankfurter and
Friendly -- has had little impact on the writing of judicial opinions, where the canons seem to be flourishing as vigor-
ously as ever. n26 This persistent gap between scholarly and practical thinking must be due, in part at least, to the lack of
systematic attention that statutory interpretation receives in the law schools. The professors drum into their students'
heads a distrust of legal formalism, and this has had an effect on opinions. Judicial opinions in America are less formal-
istic than they once were; courts are less prone to pretend that their conclusions follow by ineluctable logic from
premises found in earlier cases, without any leavening of policy or common sense. But judicial opinions continue to
pretend far more often than they should that the interpretation of statutes is the mechanical application of well under-
stood interpretive principles -- [*806] the canons -- to legislative materials.
The usual criticism of the canons, forcefully advanced by Professor Llewellyn many years ago, is that for every
canon one might bring to bear on a point there is an equal and opposite canon, n27 so that the outcome of the interpretive
process depends on the choice between paired opposites -- a choice the canons themselves do not illuminate. (You need
a canon for choosing between competing canons, and there isn't any.) I think the criticism is correct, but I also think that
most of the canons are just plain wrong, and it is that point that I want to develop here.
There is an initial question of what precisely it means to call a canon of statutory construction "wrong." The answer
depends on what a canon's function is. There are several possibilities. First, a canon might be part of a code that Con-
gress uses when it writes statutes. Suppose Congress decided that if the meaning of a statute as applied to some prob-
lem is plain as a linguistic matter, the statute should be interpreted in accordance with that meaning, even though it is
contrary to Congress's actual purpose in enacting the statute. So if Congress grants a tax exemption to "minister[s] of
the gospel," n28 rabbis should not be held eligible, n29 and if that makes the exemption unconstitutional under the first
amendment because it discriminates against a religious faith, too bad.
I do not think that any of the canons of statutory construction can be defended on the theory that they are keys to
deciphering a code. There is no evidence that members of Congress, or their assistants who do the actual drafting, know
the code or that if they know, they pay attention to it. Nor, in truth, is there any evidence that they do not; it is remark-
able how little research has been done on a question that one might have thought lawyers would regard as fundamental
to their enterprise. Probably, though, legislators do not pay attention to it, if only because, as Llewellyn showed, the
code is internally inconsistent. We should demand evidence that statutory draftsmen follow the code before we erect a
method of interpreting statutes on the improbable assumption that they do.
A second line of defense of the canons is that they, or at least some of them, are common sense guides to interpreta-
Page 4
50 U. Chi. L. Rev. 800, *

tion. It is this defense that I shall be questioning at length, by denying that the canons (with two closely related excep-
tions) have value even as [*807] flexible guideposits -- rebuttable presumptions -- rather than rigid rules. A third line
of defense is that even if the canons do not make very good sense, it is better that the judges should feel constrained by
some interpretive rules than free to roam at large in a forest of difficult interpretive questions; but I shall argue shortly
that the effect of the canons is the opposite of constraining.
There is a fourth line of defense: the canons limit the delegation of legislative power to the courts. The "plain
meaning" rule n30 forces the legislature to draft statutes carefully; the rule that repeals by implication are not favored
limits the scope of newly enacted statutes; the rule that statutes in derogation of the common law are to be construed
strictly narrows the scope of all statutes applied in areas where common law principles would otherwise govern. But of
course other canons look in the opposite direction, such as the important canon that remedial statutes are to be construed
broadly. And, as noted earlier, two inconsistent canons can usually be found for any specific question of statutory con-
struction. It is therefore unlikely that the canons considered as a whole stand for some general principle of limited gov-
ernment and separation of powers. No doubt one could, by picking and choosing, impose such a principle. But I know
of no neutral, nonpolitical basis on which a judge can decide whether the legislature should be forced by some version
of strict construction to legislate less or encouraged by some version of loose construction to legislate more. I shall
come back to this point, however, in the last part of the paper.

B. Specific Canons
I begin my discussion of specific canons with one that has both a logical priority and an apparent reasonableness
that many of the others lack. A milder version of the older, and still frequently invoked, "plain meaning" rule, n31 it
holds that in interpreting a statute you should begin, though maybe not end, with the words of the statute. n32 Offered as
a description of what judges do, the proposition is false. The judge rarely starts his inquiry with the words of [*808]
the statute, and often, if the truth be told, he does not look at the words at all. This is notoriously true with regard to the
Constitution. More often than not, briefs and judicial opinions dealing with free speech, due process, the right to assist-
ance of counsel, and other constitutional rights do not quote the language of the applicable provision -- and not because
all concerned known these provisions by heart. The constitutional provisions are in reality the foundations, or perhaps
in some cases the pretexts, for the evolution of bodies of case law that are the starting point and usually the ending point
of analysis for new cases.
There are many statutes of which this is also true. The one I know best is the Sherman Act. n33 Lawyers and judges
do not begin their analyses of a challenged practice by comparing the practice with the language of the Act and, only if
they have satisfied themselves that there is some relationship, then proceed to analyze the case law. They start with the
case law and may never return to the statutory language -- to "restrain trade or commerce" or to "attempt or conspire to
monopolize." Even in dealing with statutes that have not generated a huge body of case law, a judge usually begins not
with the language of the statute but with some conception of its subject matter and the likely purpose -- if only one de-
rived from the name of the statute or the title of the U.S. Code in which it appears. He is right to do so, because it is im-
possible to make sense of statutory language without some context.
I have thus far asumed that the "start with the words" canon has reference to temporal rather than to logical priority,
and that is I think how it is usually meant. But maybe this is being too literal and what really is intended is that the lan-
guage of a statute be deemed the most important evidence of its meaning -- which it normally is -- or at least indispens-
able evidence -- which it always is. It is ironic that a principle designed to clarify should be so ambiguous. Of course
the words of a statute are always relevant, often decisive, and usually the most important evidence of what the statute
was meant to accomplish. I merely object to the proposition that one must always begin with the words, and I am reas-
onably confident that more often than not the judge -- the good judge as well as the bad judge -- in fact begins some-
where else.
The "start with the words" canon, like the "plain meaning" canon itself, goes wrong by being unrealistic about how
judges read statutes. Another very popular canon, "remedial statutes are to be [*809] construed broadly," n34 goes
wrong by being unrealistic about elgislative objectives. The idea behind this canon is that if the legislature is trying to
remedy some ill, it would want the courts to construe the legislation to make it a more rather than a less effective rem-
edy for that ill. This would be a sound working rule if every statute -- at least every statute that could fairly be charac-
terized as "remedial" (which I suppose is every regulatory statute that does not prescribe penal sanctions and so comes
under another canon, which I discuss later) -- were passed because a majority of the legislators wanted to stamp out
some practice they considered to be an evil; presumably they would want the courts to construe the statute to advance
Page 5
50 U. Chi. L. Rev. 800, *

that objective. But if, as is often true, the statute is a compromise between one group of legislators that holds a simple
remedial objective but lacks a majority and another group that has reservations about the objective, a court that con-
strued the statute broadly woudl upset the compromise that the statute was intended to embody.
Another facet of the same point, which I have discussed elsewhere, n35 is that the absence of effective statutory rem-
edies for violations of statutory commands should not automatically be considered an invitation to judges to create such
remedies. The statute may reflect a compromise between those who wanted it to be fully effective in achieving its
stated objective and those who wanted a less effective statute; if so, it should be enforced according to that compromise.
Both the principle of supplementing weak statutory remedies with strong judicial remedies and the canon that remedial
statutes are to be read broadly ignore the role of compromise in the legislative process and, more fundamentally, the role
of interest groups, whose clashes blunt the thrust of many legislative initiatives.
The use of postenactment legislative materials to interpret a statute invites a similar objection. n36 Postenactment
statements are likely to reflect the current preferences of legislators and of the interest groups that determine or at least
influence those preferences, but the current preferences bear no necessary relationship to those of the enacting legislat-
ors, who may have been reacting to a different constellation of interest-group pressures. To give effect to the current le-
gislators' preferences is to risk spoiling the deal [*810] cut by the earlier legislators -- to risk repealing legislation, in
whole or in part, without going through the constitutionally prescribed processes for repeal. One cannot assume a con-
tinuity of view over successive Congresses.
A court should adhere to the ancting legislature's purposes (so far as those purposes can be discerned) even if it is
certain that the current legislature has different purposes and will respond by amending the relevant legislation to re-
verse the court's interpretation. The court's adherence to the initial compromise will not be futile, for the amending le-
gislation will probably be prospective (that is, applicable only to conduct taking place after the date of amendment), but
judicial interpretations of legislation are retrospective (that is, applicable to past conduct at issue in a pending case).
Thus if the court were to implement the preferences of the current legislature, it would in effect be repealing the statute
earlier than the legislature itself would have repealed it.
And all this assumes that the court can predict the preferences of the current legislature, but of course it cannot. It
is one thing to use a committee report to explain the meaning of a statute passed on the committee's recommendation; it
is another thing to rely on a committee's report that did not result in legislation to predict how the entire legislature will
act if the court does not interpret the existing statute in a particular way. Judges cannot make such predictions with any
confidence.
I do not want totally to anathematize the use of postenactment materials to interpret a statute, for such materials
may in some cases reflect a disinterested and informed view by a committee that is monitoring the administration of a
statute; and I also want to distinguish sharply between postenactment materials and a subsequently enacted statute. Ob-
viously a statute can change the meaning of an earlier statute even if the later statute does not expressly amend the earli-
er; I shall have something to say in a moment about the canon against implied repeals. But a committee report or a
statement on the floor cannot amend an enacted statute, implicitly or explicitly, and rarely will it cast much light on the
meaning of the statute.
Another canon that rests on an unrealistic view of the political process is the canon that the interpretation of a stat-
ute by the administrative agency that enforces it is entitled to great weight by the courts. n37 There is no reason to expect
administrative agency [*811] members, appointed and confirmed long after the enactment of the legislation they are
enforcing, to display a special fidelity to the original intent of the legislation rather than to the current policies of the
Administration and the Congress. They may of course know more than the courts about the legislation, and to the ex-
tent they support their interpretation with reasons at least plausibly based on superior knowledge the courts should give
that interpretation weight. But the mere fact that it is the current agency interpretation does not entitle it to any particu-
lar weight. If the interpretation has persisted through several changes of Administration, that may be a different matter.
Most canons of statutory construction go wrong not because they misconceive the nature of judicial interpretation
or of the legislative or political process but because they impute omniscience to Congress. Omniscience is always an
unrealistic assumption, and particularly so when one is dealing with the legislative process. The basic reason why stat-
utes are so frequently ambiguous in application is not that they are poorly drafted -- though many are -- and not that the
legislators failed to agree on just what they wanted to accomplish in the statute -- though often they do fail -- but that a
statute necessarily is drafted in advance of, and with imperfect appreciation for the problems that will be encountered in,
its application. All this has been explained by Edward Levi in words that I cannot improve on. He points out that the
Page 6
50 U. Chi. L. Rev. 800, *

ambiguity of a statute in application -- the incompleteness of the statute --


is not the result of inadequate draftsmanship, as is so frequently urged. Matters are not decided until they have to
be. For a legislature perhaps the pressures are such that a bill has to be passed dealing with a certain subject. But the
precise effect of the bill is not something upon which the members have to reach agreement. If the legislature were a
court, it would not decide the precise effect until a specific fact situation arose demanding an answer. Its first pro-
nouncement would not be expected to fill in the gaps. But since it is not a court, this is even more true. It will not be
required to make the determination in any event, but can wait for the court to do so. There is a related and an additional
reason for ambiguity. As to what type of situation is the legislature to make a decision? Despite much gospel to the
contrary, a legislature is not a fact-finding body. There is no mechanism, as there is with a court, to require the legis-
lature to sift facts and to make a decision about specific situations. There need be no agreement about what the situation
is. The members of the [*812] legislative body will be talking about different things; they cannot force each other to
accept even a hypothetical set of facts. The result is that even in a non-controversial atmosphere just exactly what has
been decided will not be clear. n38
An example of a canon founded on the assumption of legislative omniscience is the canon that every word of a stat-
ute must be given significance; nothing in the statute can be treated as surplusage. n39 No one would suggest that judi-
cial opinions or academic articles contain no surplusage; are these documents less carefully prepared than statutes?
There is no evidence for this improbable proposition; what evidence we have, much of it from the statutes themselves, is
to the contrary. True, statutory language is in an important sense more compact than the language of judicial opinions
and law-review articles. Every word in a statute counts -- every word is a constitutive act -- whereas much in a judicial
opinion will merely be explanatory of its holding, and much in an academic article merely explanatory of its thesis or
findings. But it does not follow that statutes are more carefully drafted, or even that greater care assures greater eco-
nomy of language; a statute that is the product of compromise may contain redundant language as a by-product of the
strains of the negotiating process.
Consider now the popular canon that repeals by implication are not favored, n40 and imagine what the idea behind it
might be. Maybe it is that whenever Congress enacts a new statute it combs the United States Code for possible incon-
sistencies with the new statute, and when it spots one, it repeals the inconsistency explicitly. But this would imply legis-
lative omniscience in a particularly uncompromising and clearly unrealistic form, for if Congress could foresee every
possible application of a new statute and make provision for it, there would be no need for judicial interpretation at all.
Since total foresight is not possible, if some latent inconsistency becomes actual all a court can do is figure out as best it
can whether Congress would have wanted to forbid the inconsistent application of the old statute or give less scope to
the new one.
An alternative basis for this canon is the idea that if the choice is between giving less scope to the new statute and
cutting [*813] down the intended scope of the old (because both cannot be enforced fully without conflict), Congress
must desire the courts to do the first. But there is no basis for this imputation of congressional purpose, and the opposite
inference is if anything more plausible -- that the enacting Congress cares more about its statutes than those of previous
Congresses.
The canon expressio unius est exclusio alterius n41 is also based on the assumption of legislative omniscience, be-
cause it would make sense only if all omissions in legislative drafting were deliberate. It seemed dead for a while, n42
but it was resurrected by the Supreme Court a few years ago to provide a basis for refusing to create private remedies
for certain statutory violations. n43 Its very recent disparagement by a unanimous Court puts its future in doubt n44 -- or
maybe just shows that judicial use of the canons of construction is hopelessly opportunistic. Whether the result in the
private-action cases is right or wrong, the use of expressio unius is not helpful. If a statute fails to include effective rem-
edies because the opponents were strong enough to prevent their inclusion, the courts should honor the legislative com-
promise. But if the omission was an oversight, or if Congress thought the courts would provide appropriate remedies
for statutory violations as a matter of course, the judges should create the remedies necessary to carry out the legis-
lature's objectives:
[t]he major premise of the conclusion expressed in a statute, the change of policy that induces the enactment, may
not be set out in terms, but it is not an adequate discharge of duty for courts to say: We see what you are driving at, but
you have not said it, and therefore we shall go on as before. n45
My last example of a canon apparently premised on an assumption of legislative omniscience is one that even
Judge Friendly, our most trenchant living critic of the canons of statutory construction, has occasionally, though cau-
Page 7
50 U. Chi. L. Rev. 800, *

tiously, invoked: that the reenactment without change of a statute that the courts have interpreted [*814] in a particular
way may be taken as evidence that the reenactment adopts that construction. Consider Judge Friendly's example of the
domestic-relations exception to the diversity jurisdiction of the federal courts. n46 This entirely judge-made exception,
although uncertain in scope, is almost as old as the federal courts themselves. n47 The grant of diversity jurisdiction to
the federal courts has been reenacted several times since the exception was first recognized, n48 yet neither the text nor
legislative history of the successive reenactments has ever referred to it. Can we nevertheless take these reenactments,
or at least the most recent, as signifying legislative adoption of the judicially created exception? Probably not. It seems
as likely that a majority of the legislators who voted on each reenactment never heard of the exception, which is in fact
unknown to all but a small number of specialists in federal jurisdiction and domestic relations, or that they heard of it
but had no desire to freeze the existing judicial construction into statute law, being indifferent to whether the courts con-
tinued to recognize the exception or decided to abolish it. n49
I could go on denouncing the canons of statutory construction, but I have discussed the ones that appear most fre-
quently today in judicial decisions and I want to turn now to three canons that have some arguable merit. I have dis-
cussed the first of these canons -- the canon that penal statutes should be construed narrowly n50 -- elsewhere. n51 Here I
add only that this canon is bound up with the broader issue of fair notice of potential criminal liability, so that a refusal
to interpret criminal statutes narrowly could violate the familiar canon that statutes should, wherever possible, be so in-
terpreted as to be constitutional. n52 This canon rests on the [*815] common-sence assumption that the legislators
would rather not have the courts nullify their effort entirely unless the interpretation necessary to save it would pervert
the goals of the legislature in enacting it. It is the cy pres doctrine applied to legislation and provides reason enough for
interpreting criminal statutes narrowly if, interpreted broadly, they would violate due process.
The next canon is related. It is that statutes should be construed not only to save them from being invalidated but to
avoid even raising serious constitutional questions. n53 Judge Friendly has criticized this canon with his customary
power. n54 He asks why the legislature should care that its statute raises a constitutional question, so long as the court
concludes that it is constitutional. n55 If the court is inclined to hold the statute unconstitutional, then the previous canon
on construing to avoid unconstitutionality, which Judge Friendly accepts, n56 comes into play. This criticism is convin-
cing as far as it goes but, as Judge Friendly recognizes, it is incomplete. n57 It leaves out of account the policy -- derived
from the structure of the Constitution -- of avoiding unnecessary constitutional decisions. Applying the canon that con-
stitutional questions are to be avoided wherever possible leaves everything pleasantly vague. Congress can amend the
statute if it feels strongly and so precipitate a constitutional controversy that it may lose (not that it must lose, as would
be the case if it amended a statute to nullify a construction that was necessary to make the statute constitutional), but if it
does not amend the statute a collision with the courts has been averted. And even if the courts were to uphold the stat-
ute's constitutionality if forced to grasp the nettle, in the course of doing so they might say something that would cast a
constitutional shadow on some other legislation. Construing legislation to avoid constitutional questions, as well as to
avoid actual nullification, is thus one of those buffering devices, much discussed by the late Alexander Bickel, by which
the frictions created by the institution of judicial review are minimized. n58
[*816] But this does not prove that the canon is a good one. It just shifts the plane of analysis from that of inter-
preting legislative intent to that of maintaining a proper separation of powers. And I think on this other plane it flops
too, so that in the end I agree with Judge Friendly. The Constitution as interpreted in modern cases is extraordinarily
far-reaching -- a written Constitution in name only. Congress's practical ability to overrule a judicial decision miscon-
struing one of its statutes, given all the other matters pressing for its attention, is less today than ever before, and prob-
ably was never very great. The practical effect of interpreting statutes to avoid raising constitutional questions is there-
fore to enlarge the already vast reach of constitutional prohibition beyond even the most extravagant modern interpreta-
tion of the Constitution -- to create a judge-made constitutional "penumbra" that has much the same prohibitory effect as
the judge-made (or at least judge-amplified) Constitution itself. And we do not need that.
If I am right that the canon of narrow construction for penal statutes is really just an aspect of the canon that statutes
should be construed to avoid being held unconstitutional, then I am down to just one canon; the rest, I respectfully sug-
gest, should be discarded. But where does this leave us? Might it not be better to subject the judges to the discipline of
the canons, even if the canons are in some ultimate sense wrong, than to invite them to approach the task of statutory
construction without any standards at all to guide them? I take up that question next. But before doing so I want to
raise the question whether the canons, far from imposing a discipline of any sort on judges, do not have the opposite ef-
fect -- promoting "judicial activism," in the sense of an expansive approach to the power of courts vis-a-vis the other
branches of government. Vacuous and inconsistent as they mostly are, the canons do not constrain judicial decision
making but they do enable a judge to create the appearance that his decisions are constrained. A standard defense of ju-
Page 8
50 U. Chi. L. Rev. 800, *

dicial activism, in the words of a defender, is that it "is, in most instances, not activism at all. Courts do not relish mak-
ing such hard decisions and certainly do not encourage litigation on social or political problems. But... the federal judi-
ciary... has the paramount and the continuing duty to uphold the law." n59 By making statutory interpretation seem mech-
anical rather than creative, the canons conceal, often from the reader of the judicial opinion and sometimes from the
writer, the extent to [*817] which the judge is making new law in the guise of interpreting a statute or a constitutional
provision. You will find more skepticism about the canons of construction in the opinions of practitioners of judicial
self-restraint than in the opinions of judicial activists. The judge who recognizes the degree to which he is free rather
than constrained in the interpretation of statutes, and who refuses to make a pretense of constraint by parading the can-
ons of construction in his opinions, is less likely to act wilfully than the judge who either mistakes freedom for con-
straint or has no compunctions about misrepresenting his will as that of the Congress.
III. AN ALTERNATIVE TO THE CANONS
I offer not a substitute algorithm but only an attitude, or maybe a slogan, and leave it to the reader to choose
between what seems to me to be the delusive rigor of the canons and the guidance offered by my suggested approach. I
suggest that the task for the judge called upon to interpret a statute is best described as one of imaginative reconstruc-
tion. n60 The judge should try to think his way as best he can into the minds of the enacting legislators and imagine how
they would have wanted the statute applied to the case at bar.
Now it is easy to ridicule this approach by saying that judges do not have the requisite imagination and that what
they will do in practice is assume that the legislators were people just like themselves, so that statutory construction will
consist of the judge's voting his own preferences and ascribing them to the statute's draftsmen. But the irresponsible
judge will twist any approach to yield the outcomes that he desires and the stupid judge will do the same thing uncon-
sciously. If you assume a judge who will try with the aid of a reasonable intelligence to put himself in the place of the
enacting legislators, then I believe he will do better if he follows my suggested approach than if he tries to apply the
canons.
[*818] The judge who follows this approach will be looking at the usual things that the intelligent literature on
statutory construction tells him to look at -- such as the language and apparent purpose of the statute, its background and
structure, its legislative history (especially the committee reports and the floor statements of the sponsors), and the bear-
ing of related statutes. But he will also be looking at two slightly less obvious factors. One is the values and attitudes,
so far as they are known today, of the period in which the legislation was enacted. It would be foolish to ascribe to le-
gislators of the 1930's or the 1960's and early 1970's the skepticism regarding the size of government and the efficacy of
regulation that is widespread today, or to impute to the Congress of the 1920's the current conception of conflicts of in-
terest. It is not the judge's job to keep a statute up to date in the sense of making it reflect contemporary values; it is his
job to imagine as best he can how the legislators who enacted the statute would have wanted it applied to situations that
they did not foresee.
Second, and in some tension with the first point, the judge will be alert to any sign of legislative intent regarding
the freedom with which he should exercise his interpretive function. Sometimes a statute will state whether it is to be
broadly or narrowly construed; n61 more often the structure and language of the statute will supply a clue. If the legis-
lature enacts into statute law a common law concept, as Congress did when it forbade agreements in "restraint of trade"
in the Sherman Act, that is a clue that the courts are to interpret the statute with the freedom with which they would con-
strue and apply a common law principle -- in which event the values of the framers may not be controlling after all.
The opposite extreme is a statute that sets out its requirements with some specificity, especially against a back-
ground of dissatisfaction with judicial handling of the same subject under a previous statute or the common law (much
federal labor and regulatory legislation is of this character). Here it is probable that the legislature does not want the
courts to paint with a broad brush in adapting the legislation to the unforeseeable future. The Constitution contains sev-
eral such provisions -- for example, the provision that the President must be thirty-five years old. This provision does
not invite construction; it does not invite a court to recast the provision so that it reads, "the President must be either
thirty-five or mature." There is nothing the court could point to [*819] that would justify such an interpretation as con-
sistent with the framers' intent. It is not that the words are plain; it is that the words, read in context as words must al-
ways be read in order to yield meaning, do not authorize any interpretation except the obvious one.
The approach I have sketched -- a word used advisedly -- in this part of the paper has obvious affinities with the
"attribution of purpose" approach of Hart and Sacks, n62 the antecedents of which go back almost 400 years. n63 But I
should like to stress one difference between my approach and thirs. They say that in construing a statute a court "should
Page 9
50 U. Chi. L. Rev. 800, *

assume, unless the contrary unmistakably appears, that the legislature was made up of reasonable persons pursuing reas-
onable purposes reasonably." n64 Coupled with their earlier statement that in trying to divine the legislative will the court
should ignore "short-run currents of political expedience," n65 Hart and Sacks appear to be suggesting that the judge
should ignore interest groups, popular ignorance and prejudices, and other things that deflect legislators from the single-
minded pursuit of the public interest as the judge would conceive it. But to ignore these things runs the risk of attribut-
ing to legislation not the purposes reasonably inferable from the legislation itself, but the judge's own conceptions of the
public interest. Hart and Sacks were writing in the wake of the New Deal, when the legislative process was widely re-
garded as progressive and public spirited. There is less agreement today that the motives behind most legislation are be-
nign. That should be of no significance to the judge except to make him wary about too easily assuming a congruence
between his concept of the public interest and the latent purposes of the statutes he is called on to interpret. He must not
automatically assume that the legislators had the same purpose that he thinks he would have had if he had been in their
shoes.
A related characteristic of the passages I have quoted from Hart and Sacks is a reluctance to recognize that many
statutes are the product of compromise between opposing groups and that a compromise is quite likely not to embody a
single consistent purpose. Of course there are difficulties for the judge, limited as he is to the formal materials of the le-
gislative process -- the statutory text, committee reports, hearings, floor debates, earlier bills, and so [*820] forth -- in
identifying compromise. A court should not just assume that a statute's apparent purpose is not its real purpose. But
where the lines of compromise are discernible, the judge's duty is to follow them, n66 to implement not the purposes of
one group of legislators, but the compromise itself. n67
But what if the lines of compromise are not clear? More fundamentally, what if the judge's scrupulous search for
the legislative will turns up nothing? There are of course such cases, and they have to be decided some way. It is inev-
itable, and therefore legitimate, for the judge in such a case to be moved by considerations that cannot be referred back
to legislative purpose. These might be considerations of judicial administrability -- what interpretation of the statute
will provide greater predictability, require less judicial factfinding, and otherwise reduce the cost and frequency of litig-
ation under the statute -- or considerations drawn from some broadly based conception of the public interest. It is al-
ways possible, of course, to refer these considerations back to Congress -- to say that Congress would have wanted the
courts, in cases where they could not figure out what interpretation would advance the substantive objectives of the stat-
ute, to adopt the "better" one, or to say a la Hart and Sacks that congressmen ought to be presumed reasonable until
shown otherwise. But these methods of imputing congressional intent are artificial; and as I argued earlier, it is not
healthy for the judge to conceal from himself that he is being creative when he is, as sometimes he has to be even when
applying statutes.
I want to end by contrasting my suggested approach with the positions in the contemporary debate over interpreta-
tion, a debate I have thus far ignored. The debate is mostly over constitutional rather than statutory interpretation, but
Professor Calabresi's recent book carries it into the statutory arena. n68 He argues that courts ought to be given, and
maybe ought to take without being given, the power to update statutes; he flirts with judicial "misreading" of statutes as
a second-best route to this end, n69 remarks [*821] that "[t]he limits of honest interpretation are too constricting," n70
and expresses at least qualified approval of judicial amendment of statutes where legislative amendment is blocked by
interest-group pressures. n71
Professor Calabresi has done us a service by bringing out into the open what are after all the secret thoughts not
only of many modern legal academics but of some modern judges. Since one extreme begets another he has also helped
us understand why there is today a revival of "strict constructionism," a revival I alluded to earlier in remarking the Su-
preme Court's recent disinterment of expressio unius. n72 But contrary to a widespread impression, strict -- that is, nar-
row -- construction, if perhaps a useful antidote to the school of no construction, is not a formula for ensuring fidelity to
legislative intent. It is almost the opposite. It is the lineal descendant of the canon that statutes in derogation of the
common law are to be strictly construed and, like that canon, was used in nineteenth-century England to emasculate so-
cial welfare legislation. n73
To construe a statute strictly is to limit its scope and its life span -- to make Congress work twice as hard to produce
the same effect. n74 The letter killeth but the spirit giveth life.
There is a story of a Vermont justice of the peace before whom a suit was brought by one farmer against another for
breaking a churn. The justice took time to consider, and then said that he had looked through the statutes and could find
nothing about churns, and gave judgment to the defendant. n75
Page 10
50 U. Chi. L. Rev. 800, *

[*822] It is not an accident that most "loose constructionists" are political liberals and most "strict constructionists" are
political conservatives. The former think that modern legislation does not go far enough, the latter that it goes too far.
Each school has developed interpretive techniques appropriate to its political ends. But as I said earlier, I know of no
principled, nonpolitical basis for a court to adopt the view that Congress is legislating too much and ought therefore to
be reined in by having its statutes construed strictly. I add now that such a view would be a form of judicial activism
because it would cut down the power of the legislative branch; and at this moment in history, we do not need more judi-
cial activism.

Legal Topics:

For related research and practice materials, see the following legal topics:
Civil ProcedureJudicial OfficersGeneral OverviewCriminal Law & ProcedureCounselRight to CounselGeneral Over-
viewGovernmentsLegislationInterpretation

FOOTNOTES:

n1 Posner, Economics, Politics, and the Reading of Statutes and the Constitution, 49 U. CHI. L. REV. 263 (1982).

n2 Landis, Statutes and the Sources of Law, in HARVARD LEGAL ESSAYS 213, 234 (R. Pound ed. 1934).

n3 See O.W. HOLMES, THE COMMON LAW (1881).

n4 See, e.g., Stigler, The Theory of Economic Regulation, 2 BELL J. ECON. & MGMT. SCI. 3 (1971); G. Becker, Competition Among
Pressure Groups for Political Influence (Univ. Chicago, Center for Study of the Economy and the State, Wkg. Paper No. 023, 1982).

n5 Posner, supra note 1, at 264, 271-72.

n6 See, e.g., K. LLEWELLYN, THE COMMON LAW TRADITION 521-35 (1960); Friendly, Mr. Justice Frankfurter and the Reading of
Statutes, in FELIX FRANKFURTER: THE JUDGE 30 (W. Mendelson ed. 1964), reprinted in H. FRIENDLY, BENCHMARKS 196 (1967);
H. Friendly, Statutorification of Federal Law (address to the Judicial Conference of the Sixth Circuit, July 1981) (on file with The University
of Chicago Law Review).

n7 See, e.g., S. BAILEY, CONGRESS MAKES A LAW (1950); R. DAVIDSON & W. OLESZEK, CONGRESS AND ITS MEMBERS
(1981); M. JEWELL & S. PATTERSON, THE LEGISLATIVE PROCESS IN THE UNITED STATES (3d ed. 1977); W. KEEFE & M.
OGUL, THE AMERICAN LEGISLATIVE PROCESS: CONGRESS AND THE STATES (5th ed. 1981); K. KOFMEHL, PROFESSIONAL
STAFFS OF CONGRESS (3d ed. 1977); M. MALBIN, UNELECTED REPRESENTATIVES: CONGRESSIONAL STAFF AND THE FU-
TURE OF REPRESENTATIVE GOVERNMENT (1980); W. OLESZEK, CONGRESSIONAL PROCEDURES AND THE POLICY PRO-
CESS (1978); N. ORNSTEIN & S. ELDER, INTEREST GROUPS, LOBBYING AND POLICYMAKING 155-85 (1978); D. PRICE, WHO
MAKES THE LAWS? (1972); R. RIPLEY, CONGRESS: PROCESS AND POLICY (2d ed. 1978).

n8 See, e.g., A. BENTLEY, THE PROCESS OF GOVERNMENT 360-81 (1908); D. TRUMAN, THE GOVERNMENTAL PROCESS 321-
94 (1951).

n9 See O. HETZEL, LEGISLATIVE LAW AND PROCESS: CASES AND MATERIALS (1980); H. LINDE, G. BUNN, F. PAFF & W.
CHURCH, LEGISLATIVE AND ADMINISTRATIVE PROCESSES (2d ed. 1981); F. NEWMAN & S. SURREY, LEGISLATION: CASES
AND MATERIALS (1955); C. NUTTING & R. DICKERSON, CASES AND MATERIALS ON LEGISLATION (5th ed. 1978); H. READ,
Page 11
50 U. Chi. L. Rev. 800, *

J. MACDONALD, J. FORDHAM & W. PIERCE, MATERIALS ON LEGISLATION (4th ed. 1981).

n10 See, e.g., O. HETZEL, supra note 9, at 693-749; H. LINDE, G. BUNN, F. PAFF & W. CHURCH, supra note 9, at 161-220; H. READ,
J. MACDONALD, J. FORDHAM & W. PIERCE, supra note 9, at 407-32.

n11 H. LINDE, G. BUNN, F. PAFF & W. CHURCH, supra note 9.

n12 Except for scattered footnote references to miscellaneous works in id. at 123-41 and a brief excerpt in C. NUTTING & R. DICKER-
SON, supra note 9, at 110-13, from Arthur Bentley's classic, The Process of Government.

n13 It is probably a forlorn hope. A recent symposium on The Law Curriculum in the 1980's, 32 J. LEGAL EDUC. 315 (1982), does not in-
clude legislation among the topics of curricular reform.

n14 On which see M. MALBIN, supra note 7.

n15 For an informative discussion of the legislative drafting process in Congress, see K. KOFMEHL, supra note 7, at 117-26, 189-93; D.
PRICE, supra note 7, and for interesting case studies, see S. BAILEY, supra note 7; N. ORNSTEIN & S. ELDER, supra note 7, at 155-85.

n16 For brief forays into this surprisingly neglected subject, see M. JEWELL & S. PATTERSON, supra note 7, at 490-93; W. KEEFE & M.
OGUL, supra note 7, at 428-32. The only extended study I have found is Note, Congressional Reversal of Supreme Court Decisions: 1945-
1957, 71 HARV. L. REV. 1324 (1958).

n17 See sources cited supra note 6.

n18 See, e.g., L. HAND, How Far is a Judge Free in Rendering a Decision?, in THE SPIRIT OF LIBERTY 103 (I. Dilliard 3d ed. 1960); 2
H. HART & A. SACKS, THE LEGAL PROCESS 1410-17 (tent. ed. 1958); R. KEETON, VENTURING TO DO JUSTICE 78-97 (1969);
Frankfurther, Some Reflections on the Reading of Statutes, 47 COLUM. L. REV. 527 (1947), reprinted in F. FRANKFURTER, OF LAW
AND MEN 44 (P. Elman ed. 1956); Friendly, supra note 6; H. Friendly, supra note 6.

n19 See, e.g., K. LLEWELLYN, supra note 6, at 521-35.

n20 Fishgold v. Sullivan Drydock & Repair Corp., 154 F.2d 785 (2d Cir.), aff'd, 328 U.S. 275 (1946).

n21 The provisions construed in Fishgold were part of the Selective Training and Service Act of 1940, ch. 720, § 8(b), 54 Stat. 885, 890 (ex-
pired 1947). These provisions were replaced by the Selective Service Act of 1948, ch. 625, § 9, 62 Stat. 604, 614 (current version at 50
U.S.C. app. § 459 (1976)).

n22 For another fine example of statutory interpretation not likely to find its way into a casebook on the statute, see J.C. Penney Co. v.
Commissioner, 312 F.2d 65 (2d Cir. 1962) (Friendly, J.).
Page 12
50 U. Chi. L. Rev. 800, *

n23 Catalogued in N. JOHNSON, SOURCES OF COMPILED LEGISLATIVE HISTORIES: A BIBLIOGRAPHY OF GOVERNMENT


DOCUMENTS, PERIODICAL ARTICLES, AND BOOKS, 1ST CONGRESS - 94TH CONGRESS (Am. Ass'n of Law Libraries, AALL
Publ. Ser. No. 14, 1979 with 1981 inserts).

n24 For a sense of this, see J. MOREHEAD, INTRODUCTION TO UNITED STATES PUBLIC DOCUMENTS (2d ed. 1978).

n25 This term has been described as a "maxim of statutory interpretation meaning that the expression of one thing is the exclusion of anoth-
er... Under this maxim, if [a] statute specifies one exception to a general rule or assumes to specify the effects of a certain provision, other
exceptions or effects are excluded." BLACK'S LAW DICTIONARY 521 (5th ed. 1979).

n26 For an amusingly dense collocation of canons, with many citations, see United States v. Scrimgeour, 636 F.2d 1019, 1022-24 (5th Cir.),
cert. denied, 454 U.S; 878 (1981). Compare United States v. Universal C.I.T. Credit Corp., 344 U.S. 218, 221 (1952) ("Generalities about
statutory construction help us little."), quoted in Weinberger v. Rossi, 456 U.S. 25, 28 (1982) with Weinberger v. Rossi, 456 U.S. 25, 32
(1982) (a principle that "has been a maxim of statutory construction since" 1804 applies in this case).

n27 E.g., K. LLEWELLYN, supra note 6, at 521-35.

n28 I.R.C. § 107 (1976).

n29 But cf. Salkov v. Commissioner, 46 T.C. 190, 193-99 (1966) (full time cantor is a "minister of the gospel" within the meaning of I.R.C.
§ 107 (1976)).

n30 A rule described as "the cardinal rule of statutory construction," requiring that "a statute's plain meaning should be given priority in its
construction." Western Union Tel. Co. v. FCC, 665 F.2d 1126, 1137 & n.21 (D.C. Cir. 1981).

n31 See id.

n32 See, e.g., International Bhd. of Teamsters v. Daniel, 439 U.S. 551, 558 (1979); Massachusetts Bonding & Ins. Co. v. United States, 352
U.S. 128, 138 (1956) (Frankfurter, J., dissenting).

n33 15 U.S.C. §§ 1-7 (1976).

n34 See, e.g., Abbott Laboratories v. Portland Retail Druggists Ass'n, 425 U.S. 1, 12 (1976).

n35 Posner, supra note 1, at 278-79.

n36 See id. at 275.

n37 See, e.g., NLRB v. Hendricks County Rural Elec. Membership Corp., 454 U.S. 170, 177, 178-90 (1981).
Page 13
50 U. Chi. L. Rev. 800, *

n38 E. LEVI, AN INTRODUCTION TO LEGAL REASONING 30-31 (1949) (footnote omitted).

n39 See, e.g., Co Petro Mktg. Group, Inc. v. Commodity Futures Trading Comm'n (In re Co Petro Mktg. Group, Inc.), 680 F.2d 566, 569-70
(9th Cir. 1982). This canon is also applied to the interpretation of contracts, see, e.g., RESTATEMENT (SECOND) OF CONTRACTS § 202
comment d (1981), just as unrealistically as it is applied to statutes.

n40 See, e.g., Kremer v. Chemical Constr. Corp., 102 S. Ct. 1883, 1890 (1982).

n41 See supra note 25.

n42 See Note, Intent, Clear Statements, and the Common Law: Statutory Interpretation in the Supreme Court, 95 HARV. L. REV. 892, 892-
96 (1982).

n43 See Touche Ross & Co. v. Redington, 442 U.S. 560, 571-72 (1979); National R.R. Passenger Corp v. National Ass'n of R.R. Passengers,
414 U.S. 453, 458 (1974). This development is described in Note, supra note 42, at 895 & n.28, and deplored in H. Friendly, supra note 6, at
10-16.

n44 Herman & MacLean v. Huddleston, 103 S. Ct. 683, 690 (1983).

n45 Johnson v. United States, 163 F. 30, 32 (1st Cir. 1908) (Holmes, J.).

n46 See Phillips, Nizer, Benjamin, Krim & Ballon v. Rosenstiel, 490 F.2d 509, 514 (2d Cir. 1973) (Friendly, J.).

n47 It had its origin in a dictum in Barber v. Barber, 62 U.S. (21 How.) 582, 584 (1859). See Phillips, Nizer, Benjamin, Krim & Ballon v.
Rosenstiel, 490 F.2d 509, 512-13 (2d Cir. 1973).

n48 The current version of the grant of diversity jurisdiction is found at 28 U.S.C. § 1332(a)(1)-(3) (1976). The previous reenactments of
that provision are as follows: Act of July 25, 1958, Pub. L. No. 85-554, § 2, 72 Stat. 415, 415; Act of June 25, 1948, ch. 646, § 1332, 62 Stat.
869, 930; Act of April 20, 1940, ch. 117, 54 Stat. 143, 143; Act of March 3, 1911, ch. 231, § 24, 36 Stat. 1087, 1091; Act of Aug. 13, 1888,
ch. 866, § 2, 25 Stat. 433, 434-35; Act of March 3, 1887, ch. 373, § 2, 24 Stat. 552, 553; Act of Dec. 1, 1873, ch. 7, § 629, 18 Stat. 1, 110-
112 (1875).

n49 A similar point is made by Judge Friendly himself in Friendly, supra note 6, at 66, reprinted in H. FRIENDLY, supra note 6, at 232-33.

n50 See, e.g., United States v. Campos-Serrano, 404 U.S. 293, 297 (1971).

n51 See Posner, supra note 1, at 280-82.


Page 14
50 U. Chi. L. Rev. 800, *

n52 See, e.g., Textile Workers' Union v. Lincoln Mills, 353 U.S. 448, 477 (1957) (Frankfurter, J., dissenting).

n53 See, e.g., United States v. Rumely, 345 U.S. 41, 45 (1953) (Frankfurter, J.).

n54 See Friendly, supra note 6, at 44-45, reprinted in H. FRIENDLY, supra note 6, at 210-12; cf. Railway Labor Executives' Ass'n v. South-
eastern Pa. Transp. Auth., 534 F. Supp. 832, 848-49 (Regional Rail Reorg. Ct. 1982) (Friendly, J.), cert. denied, 102 S. Ct. 2271 (1982).

n55 See Friendly, supra note 6, at 44, reprinted in H. FRIENDLY, supra note 6, at 210. Friendly notes that there is usually a good chance
that such questions will be resolved in favor of constitutionality. Id.

n56 Id.

n57 See id. at 44-45, reprinted in H. FRIENDLY, supra note 6, at 211-12.

n58 A. BICKEL, THE LEAST DANGEROUS BRANCH 181 (1962).

n59 Johnson, The Role of the Judiciary with Respect to the Other Branches of Government, 11 GA. L. REV. 455, 474 (1977).

n60 I associate this view primarily with Judge Learned Hand. See Lehigh Valley Coal Co. v. Yensavage, 218 F. 547, 553 (2d Cir. 1914)
(Hand, J.); L. HAND, supra note 18, at 105-10; Speech by Learned Hand, Opening Session of the National Conference on the Continuing
Education of the Bar (Dec. 16, 1958), reprinted in JOINT COMMITTEE ON CONTINUING LEGAL EDUCATION OF THE AMERICAN
LAW INSTITUTE AND THE AMERICAN BAR ASSOCATION, CONTINUING LEGAL EDUCATION FOR PROFESSIONAL COM-
PETENCE AND RESPONSIBILITY 116, 117-19 (1959), also reprinted in R. ALDISERT, THE JUDICIAL PROCESS: READINGS, MA-
TERIALS AND CASES 184, 184-85 (1976). But this view is clearly stated elsewhere, for example in J. GRAY, THE NATURE AND
SOURCES OF THE LAW 172-73 (2d ed. 1921), and has ancient antecedents, see, e.g., Heydon's Case, 76 Eng. Rep. 637, 638 (Ex. 1584); 1
W. BLACKSTONE, COMMENTARIES * 59-61; 3 id. * 430-31.

n61 See, e.g., Racketeer Influenced and Corrupt Organizations Act (RICO) § 904(a), printed at 18 U.S.C. § 1961 note (1976).

n62 See 2 H. HART & A. SACKS, supra note 18, at 1413-17.

n63 See Heydon's Case, 76 Eng. Rep. 637, 638 (Ex. 1584).

n64 2 H. HART & A. SACKS, supra note 18, at 1415.

n65 Id. at 1414.


Page 15
50 U. Chi. L. Rev. 800, *

n66 See, e.g., NLRB v. Rockaway News Supply Co., 197 F.2d 111, 115-16 (2d Cir. 1952) (Clark, J., dissenting), aff'd on other grounds, 345
U.S. 71 (1953).

n67 Compare United States v. Armour & Co., 402 U.S. 673, 681-83 (1971), where the Court interpreted the Meat Packers Consent Decree
of 1920 not with reference to the intent of the parties to the decree but rather with reference to the compromise between the parties embodied
in the decree.

n68 G. CALABRESI, A COMMON LAW FOR THE AGE OF STATUTES (1982).

n69 Id. at 34.

n70 Id. at 38.

n71 Id. at 34.

n72 See supra notes 42-43 and accompanying text.

n73 See Jones, Should Judges be Politicians? The English Experience, 57 IND. L.J. 211, 213 (1982).

n74 See Easterbrook, Statutes' Domains, 50 U. CHI. L. REV. 533, 548-49 (1983), where a theory of strict construction is explicitly defended
as a limitation on the power of legislatures to legislate with future effect.
In speaking of the effect of strict construction on the effectiveness of the legislative process, I emphasize that I am speaking only of
strict construction of statutes. If one construes the Constitution strictly, one will reduce the effectiveness of constitutional enactments; but of
course one will increase the effectiveness of the legislative process by reducing the limitations that the Constitution places on that process.
This is easily seen by thinking back to the discussion of the canon that statutes should be interpreted wherever possible to avoid being held
unconstitutional. To construe the Constitution narrowly and statutes broadly would maximize the effectiveness of that process. But the ana-
lysis of these and other interesting permutations must await another day.

n75 Holmes, The Path of the Law, 10 HARV. L. REV. 457, 474-75 (1897).